As a cloud computing technology is becoming mature, an information technology (IT) field and a communication technology (CT) field infiltrate into each other and converge together. To rise to future competition and challenges, 13 Top operators jointly issue a white paper of network functions virtualization (NFV), so as to comply with current technology development trends of virtualization and cloud computing. FIG. 1 is an NFV architecture combined with cloud computing. A network functions virtualization infrastructure (NFVI) provides a cloud-based infrastructure layer. The infrastructure layer includes hardware resources, basic software on the hardware resources, and the like that are provided by multiple manufacturers. A virtualized network function (VNF) is a cloud-based network element application. A virtualized infrastructure manager (VIM) is in charge of managing the infrastructure layer. A VNF manager (VNFM) is in charge of managing a VNF lifecycle. An NFV orchestrator (NFVO) is in charge of managing a service lifecycle. An element management system (EMS) or a network management system (NMS) is a network management system of a conventional CT telecommunication network element. An operation support system (OSS) or a business support system (BSS) is a network management system of a communications system.
The maturity of the cloud computing technology enables increasing applications to migrate to a cloud computing platform. Therefore, reliability of application services based on the cloud computing platform faces the following two challenges. On the one hand, the cloud computing platform integrates software and hardware from different manufacturers, that is, the cloud computing platform may include one or more systems, and each system includes a hardware resource and basic software that are provided by a manufacturer. In a cloud environment, software and hardware decoupling enables an application program to be deployed on an arbitrary system. Adoptable reliability mechanisms are different on different systems (a difference of the reliability mechanisms brings a reliability difference of application services on systems). Hardware resources of each system may come from different manufacturers, and their reliability is also different. Therefore, when application programs are deployed on different systems, their reliability faces a challenge. On the other hand, increasing kinds of application programs running on the cloud computing platform also bring a challenge to the reliability of the application services.
On the cloud computing platform in the prior art, reliability resource allocation is generally based on a same hardware resource at a same infrastructure layer. A reliability mechanism configured for the hardware resource is generally provided in a qualitative manner (such as a high reliability mechanism, a medium reliability mechanism, and a low reliability mechanism). An application program running at the infrastructure layer also meets a reliability requirement of a user in the qualitative manner. That is, when a client has a reliability requirement for an application service, a reliability mechanism of a corresponding level is configured for the infrastructure layer in the prior art, so as to meet the reliability requirement of the user.
However, for an industry that quantitatively requires reliability, for example, the financial industry that imposes relatively detailed reliability requirements, it is difficult for a reliability resource allocation solution of the prior art to accurately meet the reliability requirement of the user.